- Email: [email protected]
The effect of tube damage on flexural strength of ±55° angle-ply concrete-filled FRP tubes
Construction and Building Materials 240 (2020) 117948
Contents lists available at ScienceDirect
Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat
The effect of tube damage on flexural strength of ±55° angle-ply concrete-filled FRP tubes Chenxi Lu, Amir Fam ⇑ Department of Civil Engineering, Queen’s University, Kingston, ON, Canada
h i g h l i g h t s Effect of tube damage on flexural strength of concrete filled angle ply FRP tube (CFFT) is studied. Controlled longitudinal and circumferential cuts up to 22% of perimeter in length were induced. Flexural strength reduced by 64% at a hoop cut of 20% the peremeter. Circumferential cuts on tension side are more critical in cross-ply than in angle-ply CFFTs. A similar CFFT will no longer satisfy static flexural ultimate strength at a circumferential cut of 30% the perimeter.
a r t i c l e
i n f o
Article history: Received 12 September 2019 Received in revised form 18 December 2019 Accepted 23 December 2019
Keywords: Angle-ply GFRP tube CFFT Damage Cut Circumferential Longitudinal Moment reduction
a b s t r a c t The ±55° angle-ply filament-wound glass fibre reinforced polymer (GFRP) tubes are quite common commercially and have been used in concrete-filled FRP tubes (CFFTs). However, it is not clear how would a damage in the form of a cut in the tube affects the CFFT strength. This paper presents an experimental investigation to address the effects of controlled longitudinal and circumferential linear cuts, up to 22% of the perimeter (pD) in length, through the full wall thickness, on their flexural strength. The cuts were induced at mid-span on the tension and compression sides of 1420 mm long CFFTs with a 142 mm outer diameter and 4.1 mm wall thickness. Twelve specimens, including control ones without cuts to establish the full strength Mo, were tested in four-point bending. At a 10% pD cut, flexural strength dropped to 0.69Mo and 0.73Mo, respectively, for circumferential and longitudinal tension side cuts. Results were compared to another study on CFFTs with near cross-ply tubes, which were more vulnerable than angle-ply CFFTs. Compression side longitudinal and circumferential cuts in both types of tubes were less critical, with about 12% loss in strength. A design case study is also presented. Ó 2019 Elsevier Ltd. All rights reserved.
1. Introduction Concrete-filled fibre-reinforced polymer (FRP) tubes (CFFTs) have become widely researched in the past two decades. FRP tubes provide a permanent formwork and allow casting concrete directly into the tubes on site which simplifies and accelerates the construction process. The FRP strength and stiffness in each direction varies with fibre laminate structures which allows customization for individual projects. The high strength-to-weight ratio and corrosion resistance characteristics of FRP tubes make them viable alternative to steel tubes. The concrete core carries compression and provides the internal support that prevents local buckling of the outer FRP shell. ⇑ Corresponding author at: Munro Chair Professor, Queen’s University and Associate Dean Reserach. E-mail address: [email protected] (A. Fam). https://doi.org/10.1016/j.conbuildmat.2019.117948 0950-0618/Ó 2019 Elsevier Ltd. All rights reserved.
The idea of using concrete filled plastic tubes was first proposed by [1]. In the following years, researchers focused on using FRP instead of plastic tubes. The idea of filling concrete directly into prefabricated tubes was presented [2]. Unlike the internal rebar, the outer FRP shell not only increases both strength and ductility of concrete but also provides the concrete core confinement against dilation [3]. The superior concrete confinement in CFFT compression members was demonstrated by several researchers [4,5]. On the other hand, it was shown that when the specimens are subjected to bending, only partial confinement exist as small part of concrete is in compression and also because of the strain gradient [6]. To optimize the cross-section in flexure flexural behavior of CFFTs with central holes of various sizes was studied [7] and the critical concrete wall thickness was established. CFFTs have been used in field applications such as bridge components and marine piles. This includes the first bridge pier using this technology in Route 40 bridge in Virginia [8] and thousands of
2
C. Lu, A. Fam / Construction and Building Materials 240 (2020) 117948
marine piles in at least eight locations [9] with large scale experimental program to support the applications. Researchers [10] compared CFFTs with fire protection to conventional steel reinforced concrete after fire exposure. The strength of the CFFTs was not much affected and the stiffness had slightly decreased, but still performed better than steel reinforced concrete specimens. Other researchers [11] tested the effect of in-close blast loading on the behavior of CFFTs which showed that the FRP outer shell provided a protection to the inside concrete core, and the residual strength performance was better for CFFTs than counterpart steelreinforced concrete specimens. In order to enhance the sustainability of CFFTs, researchers have introduced recycled aggregate concrete and showed that 100% replacement resulted in a similar bending behavior compared to natural aggregate CFFT [13]. For pole applications, partially filled CFFTs have been studied and modeled numerically and the optimal fill height has been established [14]. This paper is focused on the effect of damage induced to the FRP tube in the form of a cut through the full FRP wall thickness, on the CFFT flexural strength for angle-ply GFRP tubes. The study is focused on the commonly available ±55° tubes used in pipeline industry and available in mass production, which would be costeffective for CFFT applications. The study explores cuts of different lengths in the circumferential and longitudinal directions and on both the tension and compression sides of the CFFT. In a previous study [12] the authors have studied the effect of cuts on CFFTs of a different type tube with near cross-ply laminate, which behaves quite differently from angle-ply tubes. This paper will also compare the two tubes in terms of their tolerance to the damage. 2. Experimental program 2.1. Test specimens and parameters Twelve CFFT specimens with an outer diameter of 142 mm and length of 1420 mm were tested under four-point bending until failure. The outer tube was ±55° angle-ply glass fibre-reinforced polymer (GFRP). The structural span was 900 mm, and the constant moment zone was 300 mm. Two intact samples F1 and F2 without any damage were used as control specimens. Damage was simulated by linear cuts made at mid-span in each of the other ten specimens. The cuts varied in length, orientation and location. F3 to F10 were specimens with cuts at extreme tension fibres. F3 to F8 had circumferential cuts with lengths of 13, 22, 45, 58, 76 and 90 mm (i.e. 3% to 20% pD). F9 and F10 had longitudinal cuts with lengths of 45 and 100 mm. F11 and F12 were specimens with cuts at the extreme compression fibres (10% and 22% pD). F11 had a 45 mm circumferential cut (10% pD), and F12 had a 71 mm longitudinal cut (16% pD), both on the compression side. Table 1 summaries the cutting details of each specimen.
2.2. Materials 2.2.1. GFRP tube All twelve CFFTs used identical glass fibre-reinforced polymer (GFRP) tubes. The outer diameter of the tube was 142 mm and the average thickness was 5.4 mm. The structural wall thickness was 4.1 mm without the resin-rich non-structural wall as indicated in the manufacture specification. The manufacturing process was filament winding method which was the most commonly used in FRP tube industry. Due to the reciprocal method of fabrication, the layers were not quite uniform along the whole specimen as well as the thickness. Burn-off test was carried out to determine the orientation of the fibre laminate. The FRP tube samples were heated to 600 °C to vaporize epoxy resin and expose the fibre structure. Three samples were tested and the laminate structure from each is reported in Table 2 and shown in Fig. 1(a). The layers of the FRP were not uniform along the whole tube, and the average angles from all three samples were [+61°/62°]. All tension and compression coupons were cut from the same FRP tube used for the CFFTs. The FRP coupons were tested in tension and compression using an Instron 8802 testing machine following a modified version of ASTM D303 and ASTM D3410. The dimensions of test coupons are shown in Fig. 1(b). For the tension tests, the dimensions of the coupons were 250 25 mm, and 50 25-mm tabs were adhered using epoxy on both sides at either end of the coupon. Two 5-mm strain gauges were installed on each tension coupon to measure the strain at mid-span on both surfaces of the coupon to establish the average strain. Three coupons were tested and the average values were recorded as the property data in Table 2. The machine was operated in a stroke control mode with a loading rate of 2 mm/min. The ultimate longitudinal tensile strength of the FRP was 54 MPa. In the elastic region, the Young’s modulus was 6.9 GPa. The ultimate tensile strain was 30,560 le. For the compression tests, three 156 25-mm coupons with 65 25-mm tabs were tested in the same machine at 1 mm/min loading rate to get the compressive strength. One 5-mm strain gauge was used to measure the strain at mid-length on each coupon. The ultimate longitudinal compressive strength of FRP was 122 MPa, and the Young’s modulus was 6.4 GPa in the elastic region. Fig. 2 shows the stress-strain curves of the tensile and compressive coupon tests. Significant nonlinear response was observed as a result of the large angles of the laminate causing shear behavior of the resin to be dominant. Fig. 1(c) shows the typical failure mode of the angle-ply FRP under tension and compression. 2.2.2. Concrete One normal strength concrete batch was used for all CFFTs. The compressive strength f’c of concrete was established from three
Table 1 Test matrix. Specimen
Cut Side
Cut Orientation
Percentage of Cut (%pD)
Length of Cut (mm)
Ultimate Applied Load (kN)
Maximum Moment due to Applied Load (kNm)
Ultimate Moment (kNm)
F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12
– – Tension
– – Circumferential
– – 3 5 10 13 17 20 10 22 10 16
– – 13 22 45 58 76 90 45 100 45 71
121.9 135.5 102.4 90.9 88.6 88.7 51.5 44.4 94.4 60.3 140.3 113.0
18.3 20.3 15.4 13.6 13.3 13.3 7.7 6.7 14.2 9.1 21.0 17.0
18.3 20.4 15.4 13.7 13.3 13.4 7.8 6.7 14.2 9.1 21.1 17.0
Longitudinal Compression
Circumferential Longitudinal
3
C. Lu, A. Fam / Construction and Building Materials 240 (2020) 117948 Table 2 Details of the FRP tube. Outer Diameter (mm) Thickness
Laminate Structure
Properties in Longitudinal Direction
Total (mm) structural wall (mm) Material Type Sample 1: [61/60/57/59/58/58/58/58/59/61/80/82/59/56/57] Sample 2: [60/59/5960/57/60/56/58/58/57/80/80/54/57/59/58] Sample 3: [57/56/57/58/55/5758/56/58/56/63/81/56/58/59/57] Average: [61°/62°] Tension E (GPa) fu (MPa) Compression E (GPa) fu (MPa)
142.4 5.4 4.1 GFRP
6.9 54 6.4 122
Fig. 1. Tension and compression coupons and laminate structure.
100 200 mm concrete cylinders tested in a Forney Testing Frame with 2224 kN loading capacity. The average concrete compressive strength was 29 MPa. 2.3. Fabrication and casting The GFRP tubes were cut using a diamond blade chop saw into 1420 mm segments before casting concrete. The hollow tubes were pinched at both ends using 8 mm diameter steel threaded rods with washers and nuts inserted radially in holes within the wall of the tube, 45 mm from the free end on either side. By tightening the nuts on both end of the threaded rod, the circular cross-section
of the tube is squeezed into an oval shape at either end of the tube. The tube recovers its round cross-sectional geometry within a short distance from either end, creating tapered ends. This taper was induced to prevent concrete slippage relative to the tube. Specimens were cast in a vertical position with the bottom ends sealed and the concrete was filled from top end. The specimens were tested after a sufficient curing period. 2.4. Simulated damage After sufficient curing of concrete, a hand-held rotary tool was used to induce controlled cuts through the entire FRP wall
4
C. Lu, A. Fam / Construction and Building Materials 240 (2020) 117948
60
2.5. Test setup and instrumentation
40
Tension
Stress (MPa)
20 0 -20 -40 -60
Compression
-80 -100 -120 -140 -80000
-60000
-40000
-20000
0
20000
40000
Strain (x 10^-6) Fig. 2. Strain-stress curve of coupon tests.
A Riehle Universal Testing Machine with loading capacity of 900 kN was used to test the CFFTs in four-point bending (Fig. 4). The tests were conducted in stroke control mode, and the loading rate varied from 1 mm/min to 2 mm/min depending on the estimated deflection. The test setup used a combination of pin supports and sliding (lubricated) plates. Strain gauges (SGs), 5 mm long, and 100-mm linear potentiometers (LPs) were used in each test, to measure strains and deflection, respectively. The layout of the locations of strain gauges on the tension and compression sides, including the vicinity of the cuts, is shown in Fig. 5 (top and bottom views refer to the compression and tension sides, respectively). Two LPs were placed at midspan on each side of the specimens. The average value was the mid-span deflection.
3. Results and discussion A summary of test results including the total ultimate moment, which includes self-weight moment of the specimen and spreader beam (0.05 kN.m) is shown in Table 1. The average ultimate moment of 19.45 kN.m of the two control specimens F1 and F2 was considered as the reference full moment capacity. Figs. 6 and 7 show the load-deflection and load-strain responses of control specimens F1 and F2, which are considerably non-linear due
Circumferenal
thickness, to simulate the damage of the CFFT. The diameter of the drilling bit was 1.1 mm which was a diamond burr nail drill, shown in Fig. 3. The cuts varied in length, orientation and location in each specimen, as shown in Table 1. All cuts were made at mid-span of the specimens manually and the depth of cuts was measured to insure full wall cuts while not digging through the concrete core. Fig. 3 shows typical circumferential and longitudinal cuts.
Fig. 3. Rotary tool and sample cuts in different directions.
Fig. 4. Test setup (dimensions in mm).
Longitudinal
C. Lu, A. Fam / Construction and Building Materials 240 (2020) 117948
5
Fig. 5. Schematic of cut configurations and instrumentation.
to the GFRP laminate structure as shown by the stress-strain curves (Fig. 2). The first load drop happened at 7 kN, indicating first cracking of concrete. The slope of the curves changes gradually at around 50 kN. The strain corresponding to this load is approximately 0.005 (Fig. 7) which corresponds to the onset of the plastic plateau of the material in tension (Fig. 2). While the end preparation of the specimens helped reduce slip significantly, compared to none-tapered CFFTs tested by others, still some slip occurred, especially in F1 specimen which had more pronounced slip than F2. This is indicated by the load drop at 104 kN and the increased
softening thereafter (Fig. 6). Both F1 and F2 specimens failed due to tension rupture of the GFRP tube (Fig. 16(a)). Failure took the for of diagonal splitting cracks along the +55/55° directions along with a physical fracture of the fibres. 3.1. CFFTs damaged on tension side 3.1.1. Effect of circumferential cuts Specimens F3 to F8 had circumferential cuts at the extreme tension fibres, which varied from 3% to 20% of the perimeter. Fig. 6
6
C. Lu, A. Fam / Construction and Building Materials 240 (2020) 117948
rate almost linearly. On the other hand the strength reduction in the angle-ply CFFTs seems to be almost gradual through out the range of cut lengths. It should be noted however that data points of the angle-ply CFFTs tested in this paper are more scatter with respect to the fitted trend than the near cross-ply CFFTs data. This may be attributed to the reciprocal filament winding pattern using a large band width of fibre roving in the angle-ply tube which may possible makes its properties dependent on location on the surface. Note that sample 1 in Table 2 was one layer less than samples 2 and 3.
Fig. 6. Load-deflection responses of control CFFTs and ones with circumferential cuts on tension side.
shows the load-deflection curves of these specimens compared with control specimens, while Fig. 8 shows their load-strain curves. Fig. 6 shows the gradual reduction in strength but insignificant reduction with stiffness. The strains of SG BOTTOM 4 for specimens F4, F5 and F6 increased with load initially and then decreased. As the GFRP cut propagated due to stress concentrations at the tip of the cut, it changes the stress flow around the cut such that it by passes the strain gauge location, hence the reduced strain. Fig. 16(b) shows a typical failure mode of a CFFT with circumferential tension cut. Similar to control specimens, failure took the form of diagonal splitting cracks along the +55/55° directions along with a physical fracture of the fibres. Fig. 9 summarizes the relation between the retention of bending moment capacity (M/Mo) and the crack length at extreme tension fibres in the circumferential direction as a percentage of the perimeter (circular markers). The data was fitted with a second order polynomial. Also shown in the same figure for comparison a similar trend established earlier but for CFFTs with a near cross-ply tube having the fibres oriented very close to the longitudinal and circumferential directions (square markers) [12]. It is clear from Fig. 9 that under the same level of damage (or percentage cut of the perimeter), the strength reduction in the angle-ply CFFTs is much smaller than that in the near cross-ply CFFTs. The strength reduction in the near cross-ply CFFTs is very sharp initially up to about 2–3% pD and then continue to reduces at a lower
3.1.2. Effect of longitudinal cuts Specimens F9 and F10 were samples with longitudinal cuts at the extreme tension fibres with lengths of 10% and 22% pD. Fig. 10 shows the load-deflection curves compared with control specimens while Fig. 11 shows the load-strain curves for the specimens. Fig. 12 demonstrates the relation between bending moment capacity retention and crack length. The figure shows almost linear trend of reduction in moment capacity with cut length. Fig. 16(c) shows a sample failure mode. Failure took the form of diagonal splitting cracks along the +55/55° directions along with a physical fracture of the fibres. For specimen F9 with 10% cut, the bending strength of the CFFT dropped 27% while for F10 with 22% cut dropped 53%. Also shown in Fig. 12 are similar results for near cross-ply CFFTs [12] for comparison. Unlike circumferential direction cuts which were remarkably different for both tubes (Fig. 9), it is clear from Fig. 12 that the reductions in moment capacity due to longitudinal cuts on the tension side appear to be similar for both tube types. This phenomenon can also be illustrated in a different way: for specimen F5 with similar cut length at tension side as F9, but in the circumferential direction, the moment capacity decreased by 31%. In other words in angle-ply tube, the circumferential cut is about 15% more critical than an equal length longitudinal cut. On the other hand, in CFFTs with near cross-ply tubes [12], the circumferential cut was 100% more critical than the equal length longitudinal cut. In near angle-ply tubes longitudinal fibre layers contribute the majority of flexural strength, hence circumferential cuts are the most critical, whereas in angle-ply all layers contribute to flexural strength and both longitudinal and circumferential cuts affect all layers. 3.2. CFFTs damaged on compression side Two specimens were cut at the extreme compression fibres. Specimen F11 had a circumferential cut of 10% pD while F12 had
Fig. 7. Load-strain responses of control specimens (some curves terminates early due to gauge malfunction).
C. Lu, A. Fam / Construction and Building Materials 240 (2020) 117948
7
Fig. 8. Load-strain responses of circumferential tension cuts at extreme fibres.
a 16% pD longitudinal cut. Fig. 13 shows the load-deflection curves compared with control specimens, while Fig. 14 shows the loadstrain curves. Fig. 15 shows the bending moment reduction with cut length for F11 and F12. For sample F11 with circumferential cut, after reaching the ultimate load, there was an 8% load drop as the concrete core slipped slightly relatively to the FRP tube. Then, the specimen carried load again with a lower stiffness until failure, which occurred at a load slightly higher, 9%, than control specimens, but within typical experimental variability. Specimen F12 failed at a load 12% lower than control specimens. And also experienced some concrete slippage. Fig. 15 also shows similar
results for near cross-ply CFFTs from literature [12] for comparison, with regard to circumferential and longitudinal cuts on compression side. No specific distinction can be concluded between the two tube types but in general the damage on the compression side is much less critical than on tension side. Specimen F11 had a similar failure mode as the control specimens (Fig. 16(d)). Failure took the form of diagonal splitting cracks along the +55/55° directions along with a physical fracture of the fibres on tension. Specimen F12 had a different failure behavior. The longitudinal cut caused the tube to be vulnerable to local buckling in compression. Cracks formed at the cut tips along fibre
8
C. Lu, A. Fam / Construction and Building Materials 240 (2020) 117948
Fig. 9. Effect of circumferential cut length at extreme tension fibres on moment capacity.
of the tube which reduces drastically at higher stress levels due to the material nonlinear behavior (Fig. 2).
4. Case study for circumferential cut at extreme tension fibres
Fig. 10. Load-deflection responses of CFFTs with longitudinal cuts on tension side.
directions due to both epoxy failure and fibre rupture (Fig. 16(e)). The tube also failed on the tension side by fracture of the fibres as the control specimens (Fig. 16(f)). It is interesting to note that in a near cross-ply CFFT with a longitudinal cut on compression side [12], the critical cut length that causes tube buckling is between 29% and 68%pD. The angle-ply tube buckled at much shorter cut of 16% pD, because likely due to the significantly lower modulus
The goal of this section is to establish the damage threshold in the form of a circumferential cut on the tension side that can be tolerated before the CFFT member is no longer deemed safe for service in real life application. The idea is that reduced moment capacity (resistance) as a result of the damage should at least be equal to the applied factored moment (demand). The demand moment is a combination of the dead and live load moments. The dead load moment is assumed equal to the moment resulting from self-weight of the CFFT specimen and spreader beam, and the live load moment will be established based on the smaller value from limiting service stress criterion and limiting deflection criterion. The limiting service stress is established based on the LRFD Guide Specifications for Design of CFFTs (2012) [15]. The maximum longitudinal tensile stress in the FRP tube under all sustained live loads should not exceed 20% of the design ultimate tensile strength (fful) to avoid creep rupture. The value of fful equals to the ultimate tensile strength (fful*) of the FRP tube multiplied by a reduction
Fig. 11. Load-strain responses of CFFTs with longitudinal cuts on tension side.
C. Lu, A. Fam / Construction and Building Materials 240 (2020) 117948
9
Fig. 15. Effect of longitudinal and circumferential cut lengths at extreme compression fibres on moment capacity. Fig. 12. Effect of longitudinal cut length at extreme tension fibres on moment capacity.
Fig. 13. Load-deflection responses of CFFTs with cuts at the compression side.
factor of 0.65 for glass fibres, which represents a reduction due to environmental effect. Therefore, the maximum design tensile stress level under service loads should not exceed 13% of fful*. In this case, the ultimate tensile strength (fful*) is the actual strength from coupon tests, which is 54.2 MPa. The limiting tensile strength is 7.0 MPa which is in the elastic region (Fig. 2). The initial Young’s modulus is 6.9 GPa and the corresponding strain is 1015 le. The
moment resulting from dead load (MDL) is 0.05 kN.m, and the corresponding strain is 5.7 le (Fig. 7) which is negligible. The moment corresponding to the 1015 le is 1.88 kN.m (Fig. 7), which is considered the service live load moment based on the stress limit criterion. The deflection limit according to the National Building Code of Canada (2015) [16] should not exceed span/360, which is 2.5 mm. The corresponding moment according to the deflection limit criterion is 2.37 kN.m (Fig. 6). The limiting live load moment (MLL) is the lower value of the two moments based on the two criteria, which is 1.88 kN.m. Considering dead load and live load only, the most conservative factored combination is Mf = 1.25MDL + 1.5MLL and is used to determine the demand moment, which is Mf = 2.89 kN.m. The reduced moment resistance of the damaged CFFT (Mr) should be at least equal to Mf = 2.89 kN.m, which is 14.2% of ultimate moment resistance of the intact CFFT (control specimens). From Fig. 9, it can be seen that the crack length corresponding to a moment ratio of 14.2% is 30% of the perimeter. This means that for a similar CFFT used in a filed application subjected to static flexure, it should be taken out of service (or repaired) if subjected to a circumferential cut longer than 30% of the perimeter on the extreme tension side. It is worth noting that for CFFTs with near cross-ply tubes [12], the equivalent circumferential cut threshold was very similar (28% of the perimeter).
Fig. 14. Load-strain responses of CFFTs with cuts on compression side.
10
C. Lu, A. Fam / Construction and Building Materials 240 (2020) 117948
Fig. 16. Failure modes.
5. Conclusions
CRediT authorship contribution statement
This paper examined the effect of different through-thickness cuts induced in the GFRP tube of CFFTs on their flexural strength. The tube was a commonly used angle-play filament wound pipe with nominal ±55° laminate (±61o actual). The cuts were induced independently on the tension and compression sides of the CFFTs, in the circumferential and longitudinal directions. The cuts also varied in length. The reduced flexural strengths are established for each type of cut and compared with those of CFFTs with near cross-ply tubes subjected to the same level of damage. The following conclusions are drawn:
Chenxi Lu: Data curation, Formal analysis, Investigation, Methodology, Validation, Visualization, Writing - original draft. Amir Fam: Conceptualization, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Writing - review & editing.
1. The most critical damage is in the form of circumferential cut on the tension side. The reduced-to-control ultimate moment ratio (M/Mo) reduces almost linearly from 1.0 to 0.36 as the cut length increases from zero to 20% of the perimeter ðpDÞ. 2. Circumferential tension cuts in angle-ply CFFTs are less critical than in near cross-ply CFFTs. In the latter, (M/Mo) drops sharply from 1.0 to 0.58 at a 2%ðpDÞ cut, then gradually to 0.25 at a 20%ðpDÞ cut. The maximum difference between angle-ply and near cross-ply CFFTs occurs at around a 10%ðpDÞ cut, where (M/Mo) ratio of the near cross-ply CFFT is half that of angleply CFFT. 3. Longitudinal cut on the tension side at mid-span causes the ratio (M/Mo) to drop linearly from 1.0 at zero cut length to 0.47 at a 22%ðpDÞ cut length. This is very similar to near cross-ply CFFTs tested previously. 4. On the compression side, circumferential cuts cause negligible reduction of flexural strength, while longitudinal cuts lead to some loss. For example (M/Mo) drops from 1.0 at zero cut length to 0.88 at a 16%ðpDÞ cut length. 5. A case study on the threshold of circumferential cut damage showed that a CFFTs similar to that tested in this study would no longer meet the static flexural ultimate strength limit state requirement once the cut length reaches or exceeds 30%ðpDÞ. In this case the member must be repaired or removed from service.
Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgment The authors wish to acknowledge the financial support of the Natural Sciences and Engineering Research Council of Canada (NSERC). Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.conbuildmat.2019.117948. References [1] C.E. Kurt, 1). Concrete filled structural plastic columns, J. Struct. Div. 104 (ST1) (1978) 55–63. [2] A. Mirmiran, M. Shahawy, A new concrete-filled hollow FRP composite column, Compos. B Eng. 27 (3–4) (1996) 263–268. [3] A. Mirmiran, M. Shahawy, Behavior of concrete columns confined by fiber composites, J. Struct. Eng. 123 (5) (1997) 583–590, https://doi.org/10.1061/ (ASCE)0733-9445(1997)123:5(583). [4] A. Mirmiran, M. Shahawy, M. Samaan, H.E. Echary, Effect of column parameters on FRP-confined concrete, J. Compos. Constr. 2 (4) (1998). [5] A.Z. Fam, S.H. Rizkalla, Behavior of axially loaded concrete-filled circular fiber reinforced polymer tubes, ACI Struct. J. 98 (3) (2001), https://doi.org/10.14359/ 10217. [6] A. Fam, S. Rizkalla, Flexural behavior of concrete-filled fiber-reinforced polymer circular tubes, J. Compos. Constr. 6 (2) (2002).
C. Lu, A. Fam / Construction and Building Materials 240 (2020) 117948 [7] Y. Qasrawi, A. Fam, Flexural load tests on new spun-cast concrete-filled fiberreinforced polymer tubular poles, ACI Struct. J. 105 (6) (2008), https://doi.org/ 10.14359/20103. [8] A. Fam, M. Pando, G. Filz, S. Rizkalla, Precast piles for route 40 bridge in Virginia using concrete filled FRP tubes, PCI J. 48–3 (2003) 32–45. [9] A. Fam, R. Greene, S. Rizkalla, Field applications of concrete-filled FRP tubes for marine piles, Int. Concr. Abstracts Portal 215 (2003) 161–180. [10] A. Echevarria, A.E. Zaghi, R. Christenson, R. Plank, Residual axial capacity comparison of CFFT and RC bridge columns after fire, Polymers 7 (5) (2015). [11] Y. Qasrawi, P.J. Heffernan, A. Fam, Performance of concrete-filled FRP tubes under field close-in blast loading, J. Compos. Constr. 19 (4) (2015) 04014067, https://doi.org/10.1061/(ASCE)CC.1943-5614.0000502.
11
[12] C. Lu, J. St. Onge, A. Fam, Damage threshold of near-cross-ply tubes used in concrete-filled FRP tubes loaded in flexure, J. Compos. Constr. 24 (2) (2020) 04019063, https://doi.org/10.1061/(ASCE)CC.1943-5614.0000990. [13] H. Lee, H. Jang, W. Chung, Effect of recycled concrete on the flexural behavior of concrete-filled FRP tubes, Int. J. Concr. Struct. Mater. 13 (1) (2019), https:// doi.org/10.1186/s40069-018-0318-8. [14] A. Fam, J.K. Son, Finite element modeling of hollow and concrete-filled fiber composite tubes in flexure: part 2 – optimization of partial filling and a design method for poles, Eng. Struct. 30 (10) (2008) 2667–2676. [15] AASHTO, LRFD Guide Specifications for Design of Concrete-Filled FRP Tubes for Flexural and Axial Members, 2012, 44p. [16] National Research Council of Canada, & Associate Committee on the National Building Code, National Building Code of Canada. 1941–1990, 2015.
Contents lists available at ScienceDirect
Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat
The effect of tube damage on flexural strength of ±55° angle-ply concrete-filled FRP tubes Chenxi Lu, Amir Fam ⇑ Department of Civil Engineering, Queen’s University, Kingston, ON, Canada
h i g h l i g h t s Effect of tube damage on flexural strength of concrete filled angle ply FRP tube (CFFT) is studied. Controlled longitudinal and circumferential cuts up to 22% of perimeter in length were induced. Flexural strength reduced by 64% at a hoop cut of 20% the peremeter. Circumferential cuts on tension side are more critical in cross-ply than in angle-ply CFFTs. A similar CFFT will no longer satisfy static flexural ultimate strength at a circumferential cut of 30% the perimeter.
a r t i c l e
i n f o
Article history: Received 12 September 2019 Received in revised form 18 December 2019 Accepted 23 December 2019
Keywords: Angle-ply GFRP tube CFFT Damage Cut Circumferential Longitudinal Moment reduction
a b s t r a c t The ±55° angle-ply filament-wound glass fibre reinforced polymer (GFRP) tubes are quite common commercially and have been used in concrete-filled FRP tubes (CFFTs). However, it is not clear how would a damage in the form of a cut in the tube affects the CFFT strength. This paper presents an experimental investigation to address the effects of controlled longitudinal and circumferential linear cuts, up to 22% of the perimeter (pD) in length, through the full wall thickness, on their flexural strength. The cuts were induced at mid-span on the tension and compression sides of 1420 mm long CFFTs with a 142 mm outer diameter and 4.1 mm wall thickness. Twelve specimens, including control ones without cuts to establish the full strength Mo, were tested in four-point bending. At a 10% pD cut, flexural strength dropped to 0.69Mo and 0.73Mo, respectively, for circumferential and longitudinal tension side cuts. Results were compared to another study on CFFTs with near cross-ply tubes, which were more vulnerable than angle-ply CFFTs. Compression side longitudinal and circumferential cuts in both types of tubes were less critical, with about 12% loss in strength. A design case study is also presented. Ó 2019 Elsevier Ltd. All rights reserved.
1. Introduction Concrete-filled fibre-reinforced polymer (FRP) tubes (CFFTs) have become widely researched in the past two decades. FRP tubes provide a permanent formwork and allow casting concrete directly into the tubes on site which simplifies and accelerates the construction process. The FRP strength and stiffness in each direction varies with fibre laminate structures which allows customization for individual projects. The high strength-to-weight ratio and corrosion resistance characteristics of FRP tubes make them viable alternative to steel tubes. The concrete core carries compression and provides the internal support that prevents local buckling of the outer FRP shell. ⇑ Corresponding author at: Munro Chair Professor, Queen’s University and Associate Dean Reserach. E-mail address: [email protected] (A. Fam). https://doi.org/10.1016/j.conbuildmat.2019.117948 0950-0618/Ó 2019 Elsevier Ltd. All rights reserved.
The idea of using concrete filled plastic tubes was first proposed by [1]. In the following years, researchers focused on using FRP instead of plastic tubes. The idea of filling concrete directly into prefabricated tubes was presented [2]. Unlike the internal rebar, the outer FRP shell not only increases both strength and ductility of concrete but also provides the concrete core confinement against dilation [3]. The superior concrete confinement in CFFT compression members was demonstrated by several researchers [4,5]. On the other hand, it was shown that when the specimens are subjected to bending, only partial confinement exist as small part of concrete is in compression and also because of the strain gradient [6]. To optimize the cross-section in flexure flexural behavior of CFFTs with central holes of various sizes was studied [7] and the critical concrete wall thickness was established. CFFTs have been used in field applications such as bridge components and marine piles. This includes the first bridge pier using this technology in Route 40 bridge in Virginia [8] and thousands of
2
C. Lu, A. Fam / Construction and Building Materials 240 (2020) 117948
marine piles in at least eight locations [9] with large scale experimental program to support the applications. Researchers [10] compared CFFTs with fire protection to conventional steel reinforced concrete after fire exposure. The strength of the CFFTs was not much affected and the stiffness had slightly decreased, but still performed better than steel reinforced concrete specimens. Other researchers [11] tested the effect of in-close blast loading on the behavior of CFFTs which showed that the FRP outer shell provided a protection to the inside concrete core, and the residual strength performance was better for CFFTs than counterpart steelreinforced concrete specimens. In order to enhance the sustainability of CFFTs, researchers have introduced recycled aggregate concrete and showed that 100% replacement resulted in a similar bending behavior compared to natural aggregate CFFT [13]. For pole applications, partially filled CFFTs have been studied and modeled numerically and the optimal fill height has been established [14]. This paper is focused on the effect of damage induced to the FRP tube in the form of a cut through the full FRP wall thickness, on the CFFT flexural strength for angle-ply GFRP tubes. The study is focused on the commonly available ±55° tubes used in pipeline industry and available in mass production, which would be costeffective for CFFT applications. The study explores cuts of different lengths in the circumferential and longitudinal directions and on both the tension and compression sides of the CFFT. In a previous study [12] the authors have studied the effect of cuts on CFFTs of a different type tube with near cross-ply laminate, which behaves quite differently from angle-ply tubes. This paper will also compare the two tubes in terms of their tolerance to the damage. 2. Experimental program 2.1. Test specimens and parameters Twelve CFFT specimens with an outer diameter of 142 mm and length of 1420 mm were tested under four-point bending until failure. The outer tube was ±55° angle-ply glass fibre-reinforced polymer (GFRP). The structural span was 900 mm, and the constant moment zone was 300 mm. Two intact samples F1 and F2 without any damage were used as control specimens. Damage was simulated by linear cuts made at mid-span in each of the other ten specimens. The cuts varied in length, orientation and location. F3 to F10 were specimens with cuts at extreme tension fibres. F3 to F8 had circumferential cuts with lengths of 13, 22, 45, 58, 76 and 90 mm (i.e. 3% to 20% pD). F9 and F10 had longitudinal cuts with lengths of 45 and 100 mm. F11 and F12 were specimens with cuts at the extreme compression fibres (10% and 22% pD). F11 had a 45 mm circumferential cut (10% pD), and F12 had a 71 mm longitudinal cut (16% pD), both on the compression side. Table 1 summaries the cutting details of each specimen.
2.2. Materials 2.2.1. GFRP tube All twelve CFFTs used identical glass fibre-reinforced polymer (GFRP) tubes. The outer diameter of the tube was 142 mm and the average thickness was 5.4 mm. The structural wall thickness was 4.1 mm without the resin-rich non-structural wall as indicated in the manufacture specification. The manufacturing process was filament winding method which was the most commonly used in FRP tube industry. Due to the reciprocal method of fabrication, the layers were not quite uniform along the whole specimen as well as the thickness. Burn-off test was carried out to determine the orientation of the fibre laminate. The FRP tube samples were heated to 600 °C to vaporize epoxy resin and expose the fibre structure. Three samples were tested and the laminate structure from each is reported in Table 2 and shown in Fig. 1(a). The layers of the FRP were not uniform along the whole tube, and the average angles from all three samples were [+61°/62°]. All tension and compression coupons were cut from the same FRP tube used for the CFFTs. The FRP coupons were tested in tension and compression using an Instron 8802 testing machine following a modified version of ASTM D303 and ASTM D3410. The dimensions of test coupons are shown in Fig. 1(b). For the tension tests, the dimensions of the coupons were 250 25 mm, and 50 25-mm tabs were adhered using epoxy on both sides at either end of the coupon. Two 5-mm strain gauges were installed on each tension coupon to measure the strain at mid-span on both surfaces of the coupon to establish the average strain. Three coupons were tested and the average values were recorded as the property data in Table 2. The machine was operated in a stroke control mode with a loading rate of 2 mm/min. The ultimate longitudinal tensile strength of the FRP was 54 MPa. In the elastic region, the Young’s modulus was 6.9 GPa. The ultimate tensile strain was 30,560 le. For the compression tests, three 156 25-mm coupons with 65 25-mm tabs were tested in the same machine at 1 mm/min loading rate to get the compressive strength. One 5-mm strain gauge was used to measure the strain at mid-length on each coupon. The ultimate longitudinal compressive strength of FRP was 122 MPa, and the Young’s modulus was 6.4 GPa in the elastic region. Fig. 2 shows the stress-strain curves of the tensile and compressive coupon tests. Significant nonlinear response was observed as a result of the large angles of the laminate causing shear behavior of the resin to be dominant. Fig. 1(c) shows the typical failure mode of the angle-ply FRP under tension and compression. 2.2.2. Concrete One normal strength concrete batch was used for all CFFTs. The compressive strength f’c of concrete was established from three
Table 1 Test matrix. Specimen
Cut Side
Cut Orientation
Percentage of Cut (%pD)
Length of Cut (mm)
Ultimate Applied Load (kN)
Maximum Moment due to Applied Load (kNm)
Ultimate Moment (kNm)
F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12
– – Tension
– – Circumferential
– – 3 5 10 13 17 20 10 22 10 16
– – 13 22 45 58 76 90 45 100 45 71
121.9 135.5 102.4 90.9 88.6 88.7 51.5 44.4 94.4 60.3 140.3 113.0
18.3 20.3 15.4 13.6 13.3 13.3 7.7 6.7 14.2 9.1 21.0 17.0
18.3 20.4 15.4 13.7 13.3 13.4 7.8 6.7 14.2 9.1 21.1 17.0
Longitudinal Compression
Circumferential Longitudinal
3
C. Lu, A. Fam / Construction and Building Materials 240 (2020) 117948 Table 2 Details of the FRP tube. Outer Diameter (mm) Thickness
Laminate Structure
Properties in Longitudinal Direction
Total (mm) structural wall (mm) Material Type Sample 1: [61/60/57/59/58/58/58/58/59/61/80/82/59/56/57] Sample 2: [60/59/5960/57/60/56/58/58/57/80/80/54/57/59/58] Sample 3: [57/56/57/58/55/5758/56/58/56/63/81/56/58/59/57] Average: [61°/62°] Tension E (GPa) fu (MPa) Compression E (GPa) fu (MPa)
142.4 5.4 4.1 GFRP
6.9 54 6.4 122
Fig. 1. Tension and compression coupons and laminate structure.
100 200 mm concrete cylinders tested in a Forney Testing Frame with 2224 kN loading capacity. The average concrete compressive strength was 29 MPa. 2.3. Fabrication and casting The GFRP tubes were cut using a diamond blade chop saw into 1420 mm segments before casting concrete. The hollow tubes were pinched at both ends using 8 mm diameter steel threaded rods with washers and nuts inserted radially in holes within the wall of the tube, 45 mm from the free end on either side. By tightening the nuts on both end of the threaded rod, the circular cross-section
of the tube is squeezed into an oval shape at either end of the tube. The tube recovers its round cross-sectional geometry within a short distance from either end, creating tapered ends. This taper was induced to prevent concrete slippage relative to the tube. Specimens were cast in a vertical position with the bottom ends sealed and the concrete was filled from top end. The specimens were tested after a sufficient curing period. 2.4. Simulated damage After sufficient curing of concrete, a hand-held rotary tool was used to induce controlled cuts through the entire FRP wall
4
C. Lu, A. Fam / Construction and Building Materials 240 (2020) 117948
60
2.5. Test setup and instrumentation
40
Tension
Stress (MPa)
20 0 -20 -40 -60
Compression
-80 -100 -120 -140 -80000
-60000
-40000
-20000
0
20000
40000
Strain (x 10^-6) Fig. 2. Strain-stress curve of coupon tests.
A Riehle Universal Testing Machine with loading capacity of 900 kN was used to test the CFFTs in four-point bending (Fig. 4). The tests were conducted in stroke control mode, and the loading rate varied from 1 mm/min to 2 mm/min depending on the estimated deflection. The test setup used a combination of pin supports and sliding (lubricated) plates. Strain gauges (SGs), 5 mm long, and 100-mm linear potentiometers (LPs) were used in each test, to measure strains and deflection, respectively. The layout of the locations of strain gauges on the tension and compression sides, including the vicinity of the cuts, is shown in Fig. 5 (top and bottom views refer to the compression and tension sides, respectively). Two LPs were placed at midspan on each side of the specimens. The average value was the mid-span deflection.
3. Results and discussion A summary of test results including the total ultimate moment, which includes self-weight moment of the specimen and spreader beam (0.05 kN.m) is shown in Table 1. The average ultimate moment of 19.45 kN.m of the two control specimens F1 and F2 was considered as the reference full moment capacity. Figs. 6 and 7 show the load-deflection and load-strain responses of control specimens F1 and F2, which are considerably non-linear due
Circumferenal
thickness, to simulate the damage of the CFFT. The diameter of the drilling bit was 1.1 mm which was a diamond burr nail drill, shown in Fig. 3. The cuts varied in length, orientation and location in each specimen, as shown in Table 1. All cuts were made at mid-span of the specimens manually and the depth of cuts was measured to insure full wall cuts while not digging through the concrete core. Fig. 3 shows typical circumferential and longitudinal cuts.
Fig. 3. Rotary tool and sample cuts in different directions.
Fig. 4. Test setup (dimensions in mm).
Longitudinal
C. Lu, A. Fam / Construction and Building Materials 240 (2020) 117948
5
Fig. 5. Schematic of cut configurations and instrumentation.
to the GFRP laminate structure as shown by the stress-strain curves (Fig. 2). The first load drop happened at 7 kN, indicating first cracking of concrete. The slope of the curves changes gradually at around 50 kN. The strain corresponding to this load is approximately 0.005 (Fig. 7) which corresponds to the onset of the plastic plateau of the material in tension (Fig. 2). While the end preparation of the specimens helped reduce slip significantly, compared to none-tapered CFFTs tested by others, still some slip occurred, especially in F1 specimen which had more pronounced slip than F2. This is indicated by the load drop at 104 kN and the increased
softening thereafter (Fig. 6). Both F1 and F2 specimens failed due to tension rupture of the GFRP tube (Fig. 16(a)). Failure took the for of diagonal splitting cracks along the +55/55° directions along with a physical fracture of the fibres. 3.1. CFFTs damaged on tension side 3.1.1. Effect of circumferential cuts Specimens F3 to F8 had circumferential cuts at the extreme tension fibres, which varied from 3% to 20% of the perimeter. Fig. 6
6
C. Lu, A. Fam / Construction and Building Materials 240 (2020) 117948
rate almost linearly. On the other hand the strength reduction in the angle-ply CFFTs seems to be almost gradual through out the range of cut lengths. It should be noted however that data points of the angle-ply CFFTs tested in this paper are more scatter with respect to the fitted trend than the near cross-ply CFFTs data. This may be attributed to the reciprocal filament winding pattern using a large band width of fibre roving in the angle-ply tube which may possible makes its properties dependent on location on the surface. Note that sample 1 in Table 2 was one layer less than samples 2 and 3.
Fig. 6. Load-deflection responses of control CFFTs and ones with circumferential cuts on tension side.
shows the load-deflection curves of these specimens compared with control specimens, while Fig. 8 shows their load-strain curves. Fig. 6 shows the gradual reduction in strength but insignificant reduction with stiffness. The strains of SG BOTTOM 4 for specimens F4, F5 and F6 increased with load initially and then decreased. As the GFRP cut propagated due to stress concentrations at the tip of the cut, it changes the stress flow around the cut such that it by passes the strain gauge location, hence the reduced strain. Fig. 16(b) shows a typical failure mode of a CFFT with circumferential tension cut. Similar to control specimens, failure took the form of diagonal splitting cracks along the +55/55° directions along with a physical fracture of the fibres. Fig. 9 summarizes the relation between the retention of bending moment capacity (M/Mo) and the crack length at extreme tension fibres in the circumferential direction as a percentage of the perimeter (circular markers). The data was fitted with a second order polynomial. Also shown in the same figure for comparison a similar trend established earlier but for CFFTs with a near cross-ply tube having the fibres oriented very close to the longitudinal and circumferential directions (square markers) [12]. It is clear from Fig. 9 that under the same level of damage (or percentage cut of the perimeter), the strength reduction in the angle-ply CFFTs is much smaller than that in the near cross-ply CFFTs. The strength reduction in the near cross-ply CFFTs is very sharp initially up to about 2–3% pD and then continue to reduces at a lower
3.1.2. Effect of longitudinal cuts Specimens F9 and F10 were samples with longitudinal cuts at the extreme tension fibres with lengths of 10% and 22% pD. Fig. 10 shows the load-deflection curves compared with control specimens while Fig. 11 shows the load-strain curves for the specimens. Fig. 12 demonstrates the relation between bending moment capacity retention and crack length. The figure shows almost linear trend of reduction in moment capacity with cut length. Fig. 16(c) shows a sample failure mode. Failure took the form of diagonal splitting cracks along the +55/55° directions along with a physical fracture of the fibres. For specimen F9 with 10% cut, the bending strength of the CFFT dropped 27% while for F10 with 22% cut dropped 53%. Also shown in Fig. 12 are similar results for near cross-ply CFFTs [12] for comparison. Unlike circumferential direction cuts which were remarkably different for both tubes (Fig. 9), it is clear from Fig. 12 that the reductions in moment capacity due to longitudinal cuts on the tension side appear to be similar for both tube types. This phenomenon can also be illustrated in a different way: for specimen F5 with similar cut length at tension side as F9, but in the circumferential direction, the moment capacity decreased by 31%. In other words in angle-ply tube, the circumferential cut is about 15% more critical than an equal length longitudinal cut. On the other hand, in CFFTs with near cross-ply tubes [12], the circumferential cut was 100% more critical than the equal length longitudinal cut. In near angle-ply tubes longitudinal fibre layers contribute the majority of flexural strength, hence circumferential cuts are the most critical, whereas in angle-ply all layers contribute to flexural strength and both longitudinal and circumferential cuts affect all layers. 3.2. CFFTs damaged on compression side Two specimens were cut at the extreme compression fibres. Specimen F11 had a circumferential cut of 10% pD while F12 had
Fig. 7. Load-strain responses of control specimens (some curves terminates early due to gauge malfunction).
C. Lu, A. Fam / Construction and Building Materials 240 (2020) 117948
7
Fig. 8. Load-strain responses of circumferential tension cuts at extreme fibres.
a 16% pD longitudinal cut. Fig. 13 shows the load-deflection curves compared with control specimens, while Fig. 14 shows the loadstrain curves. Fig. 15 shows the bending moment reduction with cut length for F11 and F12. For sample F11 with circumferential cut, after reaching the ultimate load, there was an 8% load drop as the concrete core slipped slightly relatively to the FRP tube. Then, the specimen carried load again with a lower stiffness until failure, which occurred at a load slightly higher, 9%, than control specimens, but within typical experimental variability. Specimen F12 failed at a load 12% lower than control specimens. And also experienced some concrete slippage. Fig. 15 also shows similar
results for near cross-ply CFFTs from literature [12] for comparison, with regard to circumferential and longitudinal cuts on compression side. No specific distinction can be concluded between the two tube types but in general the damage on the compression side is much less critical than on tension side. Specimen F11 had a similar failure mode as the control specimens (Fig. 16(d)). Failure took the form of diagonal splitting cracks along the +55/55° directions along with a physical fracture of the fibres on tension. Specimen F12 had a different failure behavior. The longitudinal cut caused the tube to be vulnerable to local buckling in compression. Cracks formed at the cut tips along fibre
8
C. Lu, A. Fam / Construction and Building Materials 240 (2020) 117948
Fig. 9. Effect of circumferential cut length at extreme tension fibres on moment capacity.
of the tube which reduces drastically at higher stress levels due to the material nonlinear behavior (Fig. 2).
4. Case study for circumferential cut at extreme tension fibres
Fig. 10. Load-deflection responses of CFFTs with longitudinal cuts on tension side.
directions due to both epoxy failure and fibre rupture (Fig. 16(e)). The tube also failed on the tension side by fracture of the fibres as the control specimens (Fig. 16(f)). It is interesting to note that in a near cross-ply CFFT with a longitudinal cut on compression side [12], the critical cut length that causes tube buckling is between 29% and 68%pD. The angle-ply tube buckled at much shorter cut of 16% pD, because likely due to the significantly lower modulus
The goal of this section is to establish the damage threshold in the form of a circumferential cut on the tension side that can be tolerated before the CFFT member is no longer deemed safe for service in real life application. The idea is that reduced moment capacity (resistance) as a result of the damage should at least be equal to the applied factored moment (demand). The demand moment is a combination of the dead and live load moments. The dead load moment is assumed equal to the moment resulting from self-weight of the CFFT specimen and spreader beam, and the live load moment will be established based on the smaller value from limiting service stress criterion and limiting deflection criterion. The limiting service stress is established based on the LRFD Guide Specifications for Design of CFFTs (2012) [15]. The maximum longitudinal tensile stress in the FRP tube under all sustained live loads should not exceed 20% of the design ultimate tensile strength (fful) to avoid creep rupture. The value of fful equals to the ultimate tensile strength (fful*) of the FRP tube multiplied by a reduction
Fig. 11. Load-strain responses of CFFTs with longitudinal cuts on tension side.
C. Lu, A. Fam / Construction and Building Materials 240 (2020) 117948
9
Fig. 15. Effect of longitudinal and circumferential cut lengths at extreme compression fibres on moment capacity. Fig. 12. Effect of longitudinal cut length at extreme tension fibres on moment capacity.
Fig. 13. Load-deflection responses of CFFTs with cuts at the compression side.
factor of 0.65 for glass fibres, which represents a reduction due to environmental effect. Therefore, the maximum design tensile stress level under service loads should not exceed 13% of fful*. In this case, the ultimate tensile strength (fful*) is the actual strength from coupon tests, which is 54.2 MPa. The limiting tensile strength is 7.0 MPa which is in the elastic region (Fig. 2). The initial Young’s modulus is 6.9 GPa and the corresponding strain is 1015 le. The
moment resulting from dead load (MDL) is 0.05 kN.m, and the corresponding strain is 5.7 le (Fig. 7) which is negligible. The moment corresponding to the 1015 le is 1.88 kN.m (Fig. 7), which is considered the service live load moment based on the stress limit criterion. The deflection limit according to the National Building Code of Canada (2015) [16] should not exceed span/360, which is 2.5 mm. The corresponding moment according to the deflection limit criterion is 2.37 kN.m (Fig. 6). The limiting live load moment (MLL) is the lower value of the two moments based on the two criteria, which is 1.88 kN.m. Considering dead load and live load only, the most conservative factored combination is Mf = 1.25MDL + 1.5MLL and is used to determine the demand moment, which is Mf = 2.89 kN.m. The reduced moment resistance of the damaged CFFT (Mr) should be at least equal to Mf = 2.89 kN.m, which is 14.2% of ultimate moment resistance of the intact CFFT (control specimens). From Fig. 9, it can be seen that the crack length corresponding to a moment ratio of 14.2% is 30% of the perimeter. This means that for a similar CFFT used in a filed application subjected to static flexure, it should be taken out of service (or repaired) if subjected to a circumferential cut longer than 30% of the perimeter on the extreme tension side. It is worth noting that for CFFTs with near cross-ply tubes [12], the equivalent circumferential cut threshold was very similar (28% of the perimeter).
Fig. 14. Load-strain responses of CFFTs with cuts on compression side.
10
C. Lu, A. Fam / Construction and Building Materials 240 (2020) 117948
Fig. 16. Failure modes.
5. Conclusions
CRediT authorship contribution statement
This paper examined the effect of different through-thickness cuts induced in the GFRP tube of CFFTs on their flexural strength. The tube was a commonly used angle-play filament wound pipe with nominal ±55° laminate (±61o actual). The cuts were induced independently on the tension and compression sides of the CFFTs, in the circumferential and longitudinal directions. The cuts also varied in length. The reduced flexural strengths are established for each type of cut and compared with those of CFFTs with near cross-ply tubes subjected to the same level of damage. The following conclusions are drawn:
Chenxi Lu: Data curation, Formal analysis, Investigation, Methodology, Validation, Visualization, Writing - original draft. Amir Fam: Conceptualization, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Writing - review & editing.
1. The most critical damage is in the form of circumferential cut on the tension side. The reduced-to-control ultimate moment ratio (M/Mo) reduces almost linearly from 1.0 to 0.36 as the cut length increases from zero to 20% of the perimeter ðpDÞ. 2. Circumferential tension cuts in angle-ply CFFTs are less critical than in near cross-ply CFFTs. In the latter, (M/Mo) drops sharply from 1.0 to 0.58 at a 2%ðpDÞ cut, then gradually to 0.25 at a 20%ðpDÞ cut. The maximum difference between angle-ply and near cross-ply CFFTs occurs at around a 10%ðpDÞ cut, where (M/Mo) ratio of the near cross-ply CFFT is half that of angleply CFFT. 3. Longitudinal cut on the tension side at mid-span causes the ratio (M/Mo) to drop linearly from 1.0 at zero cut length to 0.47 at a 22%ðpDÞ cut length. This is very similar to near cross-ply CFFTs tested previously. 4. On the compression side, circumferential cuts cause negligible reduction of flexural strength, while longitudinal cuts lead to some loss. For example (M/Mo) drops from 1.0 at zero cut length to 0.88 at a 16%ðpDÞ cut length. 5. A case study on the threshold of circumferential cut damage showed that a CFFTs similar to that tested in this study would no longer meet the static flexural ultimate strength limit state requirement once the cut length reaches or exceeds 30%ðpDÞ. In this case the member must be repaired or removed from service.
Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgment The authors wish to acknowledge the financial support of the Natural Sciences and Engineering Research Council of Canada (NSERC). Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.conbuildmat.2019.117948. References [1] C.E. Kurt, 1). Concrete filled structural plastic columns, J. Struct. Div. 104 (ST1) (1978) 55–63. [2] A. Mirmiran, M. Shahawy, A new concrete-filled hollow FRP composite column, Compos. B Eng. 27 (3–4) (1996) 263–268. [3] A. Mirmiran, M. Shahawy, Behavior of concrete columns confined by fiber composites, J. Struct. Eng. 123 (5) (1997) 583–590, https://doi.org/10.1061/ (ASCE)0733-9445(1997)123:5(583). [4] A. Mirmiran, M. Shahawy, M. Samaan, H.E. Echary, Effect of column parameters on FRP-confined concrete, J. Compos. Constr. 2 (4) (1998). [5] A.Z. Fam, S.H. Rizkalla, Behavior of axially loaded concrete-filled circular fiber reinforced polymer tubes, ACI Struct. J. 98 (3) (2001), https://doi.org/10.14359/ 10217. [6] A. Fam, S. Rizkalla, Flexural behavior of concrete-filled fiber-reinforced polymer circular tubes, J. Compos. Constr. 6 (2) (2002).
C. Lu, A. Fam / Construction and Building Materials 240 (2020) 117948 [7] Y. Qasrawi, A. Fam, Flexural load tests on new spun-cast concrete-filled fiberreinforced polymer tubular poles, ACI Struct. J. 105 (6) (2008), https://doi.org/ 10.14359/20103. [8] A. Fam, M. Pando, G. Filz, S. Rizkalla, Precast piles for route 40 bridge in Virginia using concrete filled FRP tubes, PCI J. 48–3 (2003) 32–45. [9] A. Fam, R. Greene, S. Rizkalla, Field applications of concrete-filled FRP tubes for marine piles, Int. Concr. Abstracts Portal 215 (2003) 161–180. [10] A. Echevarria, A.E. Zaghi, R. Christenson, R. Plank, Residual axial capacity comparison of CFFT and RC bridge columns after fire, Polymers 7 (5) (2015). [11] Y. Qasrawi, P.J. Heffernan, A. Fam, Performance of concrete-filled FRP tubes under field close-in blast loading, J. Compos. Constr. 19 (4) (2015) 04014067, https://doi.org/10.1061/(ASCE)CC.1943-5614.0000502.
11
[12] C. Lu, J. St. Onge, A. Fam, Damage threshold of near-cross-ply tubes used in concrete-filled FRP tubes loaded in flexure, J. Compos. Constr. 24 (2) (2020) 04019063, https://doi.org/10.1061/(ASCE)CC.1943-5614.0000990. [13] H. Lee, H. Jang, W. Chung, Effect of recycled concrete on the flexural behavior of concrete-filled FRP tubes, Int. J. Concr. Struct. Mater. 13 (1) (2019), https:// doi.org/10.1186/s40069-018-0318-8. [14] A. Fam, J.K. Son, Finite element modeling of hollow and concrete-filled fiber composite tubes in flexure: part 2 – optimization of partial filling and a design method for poles, Eng. Struct. 30 (10) (2008) 2667–2676. [15] AASHTO, LRFD Guide Specifications for Design of Concrete-Filled FRP Tubes for Flexural and Axial Members, 2012, 44p. [16] National Research Council of Canada, & Associate Committee on the National Building Code, National Building Code of Canada. 1941–1990, 2015.